Tables or Sankey Diagrams? Investigating User Interaction with Different Representations of Simulation Parameters

TL;DR

This study compares Sankey diagrams and traditional tables for visualizing complex system parameters, finding Sankey diagrams significantly improve user comprehension and reduce cognitive load in engineering tasks.

Contribution

It demonstrates that flow-based Sankey diagrams enhance understanding of parameter dependencies over traditional tabular interfaces in configuration tasks.

Findings

Sankey diagrams reduce cognitive load by 51%.

They decrease interaction steps by 56%.

Parameter dependencies are more immediately visible.

Abstract

Understanding complex parameter dependencies is critical for effective configuration and maintenance of software systems across diverse domains - from Computer-Aided Engineering (CAE) to cloud infrastructure and database management. However, legacy tabular interfaces create a major bottleneck: engineers cannot easily comprehend how parameters relate across the system, leading to inefficient workflows, costly configuration errors, and reduced system trust - a fundamental program comprehension challenge in configuration-intensive software. This research evaluates whether interactive Sankey diagrams can improve comprehension of parameter dependencies compared to traditional spreadsheet interfaces. We employed a heuristic evaluation using the PURE method with three expert evaluators (UX design, simulation, and software development specialists) to compare a Sankey-based prototype to traditional tabular representations for core engineering tasks. Our key contribution demonstrates that flow-based parameter visualizations significantly reduce cognitive load (51% lower PURE scores) and interaction complexity (56% fewer steps) compared to traditional tables, while making parameter dependencies immediately visible rather than requiring mental reconstruction. By explicitly visualizing parameter relationships, Sankey diagrams address a core software visualization challenge: helping users comprehend complex system configurations without requiring deep tool-specific knowledge. While demonstrated through CAE software, this research contributes to program comprehension and software visualization by showing that dependency-aware visualizations can significantly improve understanding of configuration-intensive systems. The findings have implications for any software domain where comprehending complex parameter relationships is essential for effective system use and maintenance.